 Can you hear me? Good. Okay. So my PhD is interdisciplinary, so that means I'm based in between faculty of business and law and a computer science department, so this means I'm kind of a jack of several trades, but I'm not a master of any of them. So the topic I'm going to be talking about today is digital advertising, how it works, how it doesn't work, and how we could change the way that we match providers of goods and services with customers who need to buy them. So the internet has been called the nervous system of the 21st century. It's provided us incredible value, but the business model that a lot of the services we use to access it and to communicate with each other are free and they're subsidized by a business model that's based on surveillance of consumers, and that is advertising. So this is a $512 billion industry, how does it work? Well, it's a way of matching the supply of goods and services with customers who want to buy them, but originally the purpose of advertising was to get brands to be able to be trusted, and by placing an ad in a newspaper or a billboard, it was a way of saying that they're a trusted brand, that they weren't cowboys, they weren't going to rip people off. It was also about cultural imprinting, which is the idea that if you make an association between a product and a certain way of life or a certain set of values, then consumers can express themselves through buying the product. But those don't have much to do with the kind of advertising that funds the internet today. That is target advertising. And the idea with targeted advertising is you know something about an individual and you use that knowledge to send them a special message which is designed to influence them to buy something. So for those who don't know, so this is 20% of the ad industry, but it's the majority of the revenue of a lot of the businesses that provide free services online. So for those who don't know, this is a simplified version of how that industry works. On the one side you've got sellers who want to sell stuff. On the other side you've got consumers, buyers. In the middle you've got a very complex kind of web of intermediaries. And so the sellers will have marketers whose job it is to try and sell stuff. They'll go to advertisers who will create the creative content to advertise the products. They'll use something called a demand side platform which is an intermediary service that allows them to automatically pay for ad space. And they'll do that through either an ad network or sometimes an ad exchange. And the idea of an ad network is to aggregate all of the available ad space and allocate ads to the space. An ad exchange allows this to happen through a marketplace where there are auctions. And so a publisher, so this could be any website that has display advertising space, will engage with an ad network or an ad exchange through what's called a supply side platform. The supply side platform and the demand side platform have auctions with each other through the ad exchange. And the amount that each display opportunity goes for depends on how much is known about the user that's visiting that page. So there's a whole load of personal data that gets collected in order to do this targeting. And sometimes there'll be third party data brokers that will be involved as well. And the idea is that the more that's known about an individual, the more that the ad space is worth paying for. So that's how it works in theory. Does it actually work in practice? So just from a purely business perspective. It's not really clear whether it does or not. We do know that about 44% of all the ads that are served in this way are never actually in view. So that's clusters being on a page that is visible to a user for more than one second. The remainder, the remaining 56%, it's not clear whether they get seen either because they might be ignored by consumers, they might be blocked through ad blocking software. And there's other things like click fraud. So this is where fraudsters essentially cheat marketers out of their ad spend by generating fake clicks and views of adverts. And they charge them per click or per view. Then there's the question of whether the profiling is particularly accurate. I think in some cases it is, but in a lot of cases it isn't. I logged on to the ad preference manager of an ad network and looked at my own profile. They think that I am a father with two kids and that I have a pet dog. As far as I know, neither of those things are true. So any targeting that's targeted at me on that basis is unlikely to be correct. But maybe there are some cases where the adverts that reach people in this way are relevant to them. The theory is that this is good for consumers because it's more relevant adverts and therefore things that they might actually be interested in. But aside from that, there are a whole bunch of problems with this infrastructure for consumers. The first one is surveillance. The ad industry makes surveillance very valuable. It encourages it, it invests in it. I'm not going to go into detail on this, but I think there are lots of people talking at the Congress who will be able to articulate far better than me the dangers of surveillance for civil liberties. The second problem is the use of this infrastructure for distribution of malware. So why would you try and hack into someone's computer if you could just simply buy access to it? So if you can pose as a marketer, then you can use this infrastructure to deliver a personalised phishing attack through an advert. So the advert will resolve if it's clicked to a landing page where it serves malicious software. The other problem is that with all this personal data sloshing around and being bought and sold, it's very hard to prevent it from getting into the wrong hands. Where, for instance, a data broker might sell information about consumers. There was a case where this happened in a Vietnamese scanning outfit just bought a whole load of personal data from a data broker and then sold it on to others. And there's the problems with personalisation generally. This whole system is set up to personalised stuff, but what seems to be convenient personalisation to some engineers in Silicon Valley or wherever might to other groups seem more like discrimination. One person's personalisation is another's discrimination. And when you can't tell how this is being done, when there's no way to scrutinise the algorithms that are driving this, then I think we have a problem. And there's a final interesting psychological problem with personalisation, which is that if you get too much of it, then people tend to get freaked out. There's a parallel here with some research that's been done in the field of robotics back in the 70s, they found that they were looking at how people responded to robots with human features. And they found that a little bit of human features makes people respond better to robots. They prefer to interact with them, but only up to a point. After a certain point, there's something called the Uncanny Valley where the robot looks too much like a human and people get freaked out because it's not quite a human, but it's almost a human. And so there are some marketing companies that are seriously worrying about a similar effect being the case in personalisation that's done with targeted marketing. So what's the solution? I think there are better ways that we could devise of matching people who have needs and wants with the products and services that could fulfil them. So I'm going to go through a few different ideas that I think could be really useful. So the first one is inspired by how large organisations go about buying stuff. What they don't do is they don't sit around on their computers waiting for ads to be targeted at them that are relevant to the stuff they want to buy. That would be crazy. They have procurement teams, and the procurement teams have a job which is to send out requests for proposals. A request for proposal will say exactly what it is that they want. There are all sorts of criteria that they have, internal policies, maybe even environmental or ethical policies that they have. And then suppliers have to compete for that tender. And they do this in a reverse auction so they compete against each other to provide the service at a good price. So this kind of infrastructure could work for individuals as well. There's a service called Flubit, which is one example where you find something online that you want to buy but you don't like the price. You copy the URL, paste it into the service. Within a few hours you will get back offers from companies who want to provide you with that same product at a cheaper price. And they'll be competing against each other in order to deliver that offer to you. So there's a guy called Doc Sells from Harvard University's project VRM, which stands for Vendor Relationship Management. And he calls this intent casting. So rather than consumers responding to the messages that put out by marketers, it's actually marketers or suppliers of products and services that are responding to the intentions that are broadcast by individuals. But this still leaves the problem that as an individual I don't have a great deal of power in the marketplace. What gives me more power is being aggregated with lots of other people. So this is where collective purchasing comes in. Essentially the idea is that by aggregating demand for a particular good that an intermediary can give consumers much more bargaining power with suppliers. And this works particularly well in areas like energy. You have switching schemes that have been run where intermediaries like charities or local councils, in some cases even churches, will get people together and help them switch to an energy provider that can provide them with a better service at a lower price and so on. And the energy providers, even the big four energy providers in the UK were clamoring to try and offer the lowest price in order to get the customer. This isn't a new idea, it's been around for a long time. During the Industrial Revolution, workers, cooperatives, pulled their labour to increase the bargaining power that they had with employers, but they did the same thing with buying household goods. They had collective purchasing, they called them cooperative wholesale societies, and the aim was not just to do bulk purchases, but to actually organise production and have an influence on how the goods that they bought were made and so on. And so the final thing I think is worth looking at is assurance contracts. This is the technical term that economists use to describe a Kickstarter-style pledge system. The idea is that you get lots of people to pledge that if enough other people do so, they will put their money into a project, and then before you even go into production you have a proven demand. So it's a way of uncovering latent demand that might not otherwise be revealed through people's purchases. And some people have advocated moving towards this model for a lot of new things and they call this the pre-order economy. So something I think is important to notice with all these tools, these mechanisms, is that they're initiated by the buyer, they get to set the criteria, and so only when a sale actually goes through is there any direct contact between the provider and the buyer. And they can also be used to turn the tools of surveillance that are being used right now by marketers back on the providers. So there was a campaign called Move Your Money in the UK, which was looking at ethical banking and it gathered thousands of people together who wanted to change their bank to a more ethical bank. They did a whole load of research into which banks were most ethical and then on a certain date they will switch to the one which should manage to demonstrate enough that it was ethical. So there are other criteria that can be used, other than just a particular product at a particular price. So I think the missing piece here is the ability to do all these things in a decentralized way. At the moment all of the providers of these services are individual organizations, they build their own software, they have to aggregate demand, and they have to work in specific industries, and it's not a model that scales very well across the whole marketplace. So I think what's needed is a decentralized way of doing this. So the idea would be that your wants and your needs would be broadcast anonymously on a decentralized network. So let's say I want to buy a laptop and I want certain specifications, I want a certain operating system, there are some companies I don't want to deal with because I don't trust them for instance. So all this information can be broadcast to a decentralized network. From there it can be aggregated into a collective alliance of other people with similar needs and then suppliers can then crawl that network and look at all of the different demands that are being made and bid to meet those demands. And only then does my purchase become linked to my identity if I wanted to when I complete the purchase. And so I think there are, this is kind of an example of how it can work is a very rough sketch of how it might work. There are people working on doing this in a decentralized way, one of them is here, his name is Florian Kledorf, I think he's over there. So if you want to find out how his project is achieving this, it's called the Web of Needs, I'd recommend getting in touch with him. So the point of all this is that I think it's a better way of matching supply and demand. It avoids all of the problems of the digital advertising industry as it currently stands. And I think it's a better way of organizing consumer markets. So conclude, I'd like to put this in a broader context. There's a narrative of consumer empowerment that kind of stems from the enlightenment idea that we're all individuals with autonomous decision-making power. We interact with each other on the market and we make choices based on the options that are available to us. But I think this is a myth. It doesn't really work like that. When the web came along we thought that it was going to make it even better for consumers because there'd be even more information available at our fingertips and we'd have even more ability to decide. But actually it doesn't really work like that. There's so much information out there but most of the time we're being targeted with advertisements rather than going out and searching it all. And so I think what we need instead is a system that is based on transparency from the providers that's not based on surveillance of consumers. I think it's important to have a system that is not about grabbing the attention of consumers and trying to convince them to buy stuff but that actually is based on their needs and their desires and their individual criteria. And that treats them as individuals with their own set of values rather than just manipulable kind of predictable consumers. So I'm not sure if I've finished ahead of time but I'll be pleased to take your questions. First in answer to your question you finished well ahead of time so we've got 12 minutes for Q&A or discussion. And I see there is only one person on microphone one so go ahead. Okay thank you for your talk. First question you talked about advertisements that advert as a product but what about advertisements that advert as a brand like where a car manufacturer says hey we build great cars please like us. Well there's an interesting theory that that kind of advertising should actually be more valuable and the argument there is that targeting allows you to pay a very small amount to target a very specific person and so anyone can do that. You don't need to be a big company with a lot of money to invest and have been going for a long time to do that. So actually rationally you might think that a target advertisement signals that you're not necessarily a very reputable brand whereas if you buy a double page spread in a magazine and you to demonstrate that you have that power to do that then people might be more interested in it and be more willing to respect you as a company that can do that. So that's one theory. I think generally branding advertisements are a different beast and they don't necessarily require surveillance because they're normally higher profile. They're on billboards or in magazines. OK. And another thing if you actually next question was another microphone which would be number two. One question for me for me is I see a lot of advertising in the net advertising something that is next step for advertising. So for example if you have a free game on the iPad it has advertising for next free game that maybe gets some advertising for the real product. This is getting metameter meta levels. Will this ever break down completely because all these levels of advertising. I'm not sure I quite are you saying that you're in a game you get an advertisement for the next next free game that perhaps monetizing itself by selling another ad for another free game and I need to. Well I guess that wouldn't require that wouldn't require surveillance because all you all you need to know is that the person is playing the game right. So there aren't as many negative effects I think and that you've already expressed your interest in the game by playing it. So number four. Yeah I'm just more like a comment I think. So I basically I don't watch ads because I have an ad blocker and I haven't had the TV for like I don't know more than 10 years and things like that. So I really am not you know aware and then what's really weird is that sometimes people tell me that hey you know I've seen this ad and you know have a look it's really weird and I'm like whoa you know that's a new product and it's actually useful like you know it's the ad what I'm trying to say is that they're not pushing ad on me it's like it's actually valuable information which I know sounds very strange but so what I'm trying to say is that if we turn this around which is what you're trying to propose I think is that there will be a lot I mean there will be an information flow that will be missing. I mean I personally thought that that information flow is complete use this information but it turns out that I had to face the fact that it's not completely useless and so what do you say to that. So if the question is about you know without ads how do we discover new products that we weren't aware of. If it's completely based on our own initiative to go and make the request then we might miss out on important opportunities and I think that's right but I think there are other models to discover products that you might want to need. There are you know obvious to write blogs about particular things that are great and they look at all the new products that are coming out and that's an impartial source it's not coming from the seller and so I think you should there's always more reason to trust an impartial source than anything that comes from the seller. I think it's interesting to consider that. Could you just one question per person please and the next questions right behind you on microphone four as well. OK thank you. I'm thinking of some experiment which was about coffee where they asked some consumers what coffee they want and they all said they want some which roast black and strong coffee and when they tasted or tested different coffees they all wanted some milky soft coffee thing. So when companies look at that do you think they will trust the opinion of the customers when they see well they don't really know what they want. I think yeah I think it's true that as I said we don't really know what we really want. We usually make bad decisions and we usually make sub-optimal decisions. But part of that is about tools that allow me to be aware of my own inconsistencies that allow me to understand what I really want and when that differs from what I think I want. It's another thing to say that that's a justification for putting leaving all this infrastructure in the hands of of those who are trying to convince me to buy stuff especially if they if I think that I like a certain thing but my actions reveal I like something else and I might want to change that. Let's say it's I think I want to quit smoking. Right. I know that I want to quit smoking but I keep on smoking. That doesn't mean that I actually do like cigarettes. It's a conflict between my higher decision-making power and my my my base decision-making. Next one is number three. Thank you very much for the great talk. If more and more collectives come around then they'll become a big player in the network that you showed in the beginning. And I'm pretty sure that industry will come up with a smart business model and try to involve them in all their, you know, what they do. Have you thought about what they'll come up with and what we should be afraid of and be prepared for? Yeah. Yeah, I think I think that they see that and there's a positive opportunity for sellers at least. I think that a lot of the ad industry will if this were to happen a lot of the industry would die off. But I think from the seller perspective there's a huge advantage if you don't have to go and acquire consumers you don't have to spend money on advertising anymore. And that's great. I mean, if the energy switching program was what everyone used rather than just sort of 10% of the people using it then you could have an energy company that had no marketing budget but that exclusively so the prices would be lower and they'd always be winning the custom of the collectives. So I think there's a huge opportunity for businesses to use this model and therefore not have to go in for the whole digital advertising model. Next we have a question from the internet. So what do you think about consumer generated wish list approaches? Are those a good model? Yeah, that's a decent model. I know that Amazon has that. And yeah, I think that's a good idea. The problem at the moment is it's centralized. So if I have a wish list model with a particular vendor then I only get offers from that vendor. And so another service that could do that and that could match my wishes with similar people and aggregate them together. I think that'd be a good thing. Next one was on number one. So you talked about collective purchasing which is now popular for energy markets. You're probably a bit too young for it but like when we had the new economy in 99 and 2000 there was a lot of collective purchasing for consumer goods like cameras. So basically it worked. Some broker was collecting the requests and whenever they had an order for like 10 cameras they would send it out. But that magically vanished. So did you do any comparison? What went wrong there? Or why did it vanish? And will it vanish again? Or what did they do wrong in those years? I'm not sure. It's a good question. I think where it has been successful it's been in markets that are more like utilities than one-off purchases because people have to pay their energy bills or they have to use their bank year in, year out. And so if you have to do that it's kind of you're confronted with that decision so many times that at some point you think well maybe I should do this in a more efficient way. Whereas if you're buying a camera you might not have to consider how you make the decision so many times. So people might not know about them. I think it's hard for that kind of service to get a foothold. And also there's a problem of instant gratification. You can go and buy the camera straight away rather than having to wait for enough other people. But I think people pledge money on Kickstarter for stuff. 90% of the stuff on maybe more on Kickstarter doesn't get funded. But people still put their money down just in case. So I think this could grow. Next question is on microphone number two and I don't know whether we'll have enough time to answer all questions standing there but go ahead. Okay then I try to make it quick. So I think a lot about empowering consumers and you talked more about the functional and qualitative set of preference consumers may have when purchasing a product. But there's also the dimension of ethical, moral, political values and views that we also may wish to express by purchasing. And of course once you solve the problem of combining the purchasing power of consumers that's great. But I wonder if you've thought about more about really creating transparency on the supply side. So as long as marketers or companies can still create an image that we are making the world a better place by you know and that they have very human rights saving production process. So how can you really solve the issue of information asymmetry between what the consumer knows and what maybe a group of people may believe they know? I absolutely agree. And I think there are people working on that. So there's things like open product data, open supply chains, all that kind of stuff. And some companies are making efforts to do that. But I agree that there's a lot of asymmetry and it's very opaque at the moment. I think one model you could have would be okay, let's only deal with the ones that say what they do and provide evidence for it. And so the consumer might take a hit because they might miss out on other stuff that is being produced in an ethical way. But it provides an incentive for the suppliers to actually publish that information. Right now there's just not enough demand for it. But I think in future it'd be great if there was. And this campaign I mentioned, the Move Your Money campaign, they created a lot of that information themselves by going out and doing their own investigations into how the banks were using their money for investments. Next is on microphone four. I'm going to be a bit provocative. So I'm wondering whether you've seen any of those business models you described that were successful on a large scale because what you describe is intent-casting, so collecting consumer demand. It's been tried, and the role has been taken over by social networks. Group buying was tried by Groupon and became a total perversion, total push model. And assurance contracts kickstarted successful. It is because it's a destination site in itself because there's strong editorial control on what is allowed and kickstarted or not. So I mean, I'm just wondering, have you answered the question already yourself? Consumers are lazy. So there needs to, I mean, the role for intermediaries, will, we will still need the role for intermediaries. Yeah, I think that for now, yes, intermediaries need to be recognized brands in themselves. There are some in the UK that have to do a lot of, spend a lot of money marketing themselves and demonstrate to consumers that they exist and that what the value they provide. So I think for now, they're most likely to be centralized, but I hope that in the future they won't be. And I think for this model to scale more widely, I think that it does need to be decentralized. Well, unfortunately, you ran out of time. Maybe you find another room to continue this discussion. Thank you for your talk and for the Q&A.